For calculating DOF we are doing 3-1-1=1 and it says as we have estimated 1 parameter. If we estimate 2 parameters in some other case we would have subtracted 2 after we had subtracted 1 from the number of groups to find DOF. Secondly, Why are we using 0.5% value not 5% here?
Yes, to your first part. The question doesn't ask you to test at the 5% level. The solution uses 0.5% level for the testing. You can re-do the solution at the 5% level and get the same conclusion. The point made here is that as the chi-square value is so large that we can reject the null hypothesis even at 0.5% level.